Dear Group:

I really wouldn't recommend trying to reduce the sensitivity of the
Team Digital detectors as they are already on the very low end of all
DCC detectors that are on the market today. For example, their web
site says that you need no larger than a 4.7K resistor to activate
the detector.

For reliable detection it's always best to use a resistnace much
smaller than the rated value for the detector. This means with the
Team Digital you would want to go with say a 1.56K maximum for
reasonable reliability detecting a single car at a typical OS section.

At 1.56K per car, and 14.8V for your DCC, you are drawing 9.43mA per
car. Then say you have 300 cars on the layout in signal territory you
are drawing 3A from your booster just to cover the resistor wheel
sets.

Temperature rise per resistor is also a potentially harmful
condition. Again at 14.8V for the DCC and a 1.56K resistor you are
dissapating .14W per resistor. This means you need to use at least a
1/4W resistor which would still get quite hot. It also means that
your can't effectively use all the typically used surface mount
resistors such as .1W, .125W, etc.

You could go with higher value resistance, but with short blocks and
especially OS sections with a speck of dirt or wheels and/or track
you will absolutely not get reliable detection.

For quality detection it is important to always have the detector
rated at least 3 times greater sensitivity than the value of largest
resistance you plan to use for your wheelsets.

For example, the most popular wheelset resistance for all the
different manufactures of resistance wheelsets is 10K. That means you
want to seek out a detecotr with at least 30K sensitivity, minimum.
The Team Digital rates theirs at 4.7K.

Also, with high sensitivity, it is vitally important that the
detector have a built in sensitivity adjustment. Why, because you
basicaly want as high a sensitivity as you possiblly can and not have
faulse occupations. That condition varies for every block on your
system, partly due to leakage current between the rails which is
proportional to the length of your blocks.

Also, the shorter the block you need, and can achieve, greater
sensitivity. This is especially the case for short OS sections where
you want to reliability detect a single car so that you can't throw a
turnout under a train, i.e. if copying prototype practice.

Also, when looking for a detector you want both turn-on and turn-off
delays built into the detector to minimize dirty track problems. Just
put a VOM across a piece of your track and let a single car roll
along on it. Your meter will jump all over the place from infinity
down to values approaching the resistor in the car but hardly ever
actually stay steady anywhere near the resistor value.

Another thing to look for is how immune is the detector to pickup
from cross-coupling between adjacent block wiring. Although I've
never seen nor used the Team Digital circuit, it looks like from the
problems being reported, i.e. need to reduce a sensitivity that is
already low and the problems with false detection with long blocks
that the detector is very troubled at handling long blocks. This is
typical with transformer coupled detectors where the secondary side
of the transformer is not properly terminated with a low impedance.

This is suspect but also somewhat confirmed with the statement that
you need to keep resistor leads short which basically confirms that
the detectors performance is hindered by any nearby electrical noise
and especially noice being introduced by long block wiring which acts
like an antenna to a detector.

Lastly, some users will consider the 2 detectors per board to be an
advantage, just like 8 and 16 per board is often considered an
advantage. I personally consider more that one detector per plugable
card a disadvantage.

I personally enjoy the advantage provided by one simple plug-in
detector per block. Nothing can beat it in terms of easy system
debug. Have a problem with a short somewhere in you system it is a
snap with single modular pluggable detectors to not only quickly
locate the short but also to resolve the problem. It's also easy for
the owner to repair the detector, i.e. with a simple single sided
board and with only two active components per board, repair is also
easy to accomplish.

Just like quality trackwork is essential to achieving quality
operation, quality detection is fundamental to achieving quality
signaling.

In summary, although most every detector will interface OK with the
C/MRI, it doesn't mean that all detectors are created equal. In this
light, I hope that the above comments shed some light into a few of
the basic qualities to look for when seeking out a detector that will
give you good performance.

Sincerely,


Bruce Chubb, MMR